1 of 17

1
LISTENING TO THE WINDTVA Experience with
Installationand Operation of UltrasonicWind
Sensors

• Kenneth G. Wastrack
• Doyle E. Pittman
• John E. Hatmaker, PE
• L. Wayne Hamberger
• Tennessee Valley Authority

2
Background

• Problem - Obsolete Wind Sensors
• Difficulty obtaining spare parts
• (sensor, switch controller, etc)
• Routine bearing replacement
• History of mechanical failures

• Requirements - Sensor Specifications
• Comply with RG 1.23 requirements.
• Wind direction /-5 degrees.
• Wind speed /-0.5 mph.
• Less maintenance than existing sensors.

3
Vaisala 425AHUltrasonic Wind Sensor

• Sonic Wind Sensors
• All-electronic
• Minimal maintenance

• Vaisala (formerly Handar) 425AH ultrasonic wind
  sensor most suitable
• Manufacturer specifications
• Capabilities for integration
• System meets RG 1.23

4
System Performance
5
Principle of Operation

• Transmit signal along axis

• Send in reverse direction

• Calculate time differences

• Convert to WS along axis

• Repeat for other axes

• Compute WD and WS

6
Obtain Necessary Approvals
TVA did not obtain prior NRC approval.
Used 10 CFR 50.59 process and treated new
ultrasonic wind sensor as equivalent to old
wind vane/anemometer system.

• Same mounting positions.
• Same sampling intervals.
• Same data processing.
• No change from user point-of-view.

Obtained prior approval from Plant Operations
Review Committee (PORC) because of need for FSAR
change.
7
Develop Sensor Acceptance Criteria

• Wind Direction
• Orient sensor into each of 5 directions.
• 360, 30, 150, 270, and 45º
• Set wind tunnel speed to about 10 mph.
• Verify that sensor output equals desired values
  (/- 3º).

• Wind Speed
• Orient sensor to display wind direction of 45º.
• Adjust wind tunnel speed.
• 0, thresholdlt1, 1, 2, 5, 10, 20, and 30 mph
• Verify that sensor output equals desired values
  (/- 0.3 mph).

8
Perform Field Installation - I

• Ultrasonic wind sensors were installed on the
  following dates
• Watts Bar (WBN) - October 2000
• Browns Ferry (BFN) - February 2001
• Sequoyah (SQN) March 2002

Installations were conducted over two days
according to normal Limited Condition of
Operation (LCO) requirements for plant. Project
was examined beforehand by maintenance and
engineering personnel to address site-specific
issues. Installation was not complete until a
meteorologist verified data were acceptable and
were transmitted properly.
9
Perform Field Installation - II

• Despite careful preplanning, problems occurred.
• Proper survey equipment not available.
• Need to modify onsite acceptance test.
• Scheduling restrictions resulted in adverse
  weather impacts.
• Need for change in planned installation at BFN.
• Problems were addressed because individuals
  onsite during installation had authority to make
  adjustments to workplan.
• After verifying correct operation, meteorologist
  provided final acceptance of the installation.

10
Operational Experience
Sensor calibration process has changed. Level of
maintenance has decreased. Strip chart recorders
are no longer used. Exceedances of calibration
tolerances have decreased. Optical isolators
provide effective lightning protection.
11
New Problems

• Effect of noise on measurements.
• Consider noise environment in site exposure
  criteria.
• Consider ambient noise (e.g. thunder) during
  data validation.
• Configuration of sensor communications settings.
• Extensive effort to send sensors to laboratory.
• Developed methodology for in-situ configuration.
• Review/revise installation/exchange procedures.
• Margin verifier issues.

12
Data Changes - I
Differences in wind direction percentages between
the two types of sensors are generally within
expected ranges.
Watts Bar (10m)
Browns Ferry (90m)
13
Data Changes - II
Browns Ferry (90m)
Watts Bar (10m)
14
Data Changes - III
Significant changes have been noted in data
recovery percentage and number of calm wind
speeds observed.
Watts Bar (10m) 2000 2001
Diff. Data Recovery () 94.4 99.5
5.1 Mean Avg. WS (mph) 3.57 3.90
0.33 Number of Calms 210 9 -201
Watts Bar (90m) 2000 2001
Diff. Data Recovery () 94.3 99.4
5.1 Mean Avg. WS (mph) 7.58 7.42
-0.16 Number of Calms 4 0 -4
15
Other Tidbits - I
Plan for final installation was to move sensor to
91-meter level.
Comparison tests demonstrated that data from
93-meter and 90-meter levels are equivalent.
Moving sensor as planned was a major project.
Therefore, sensor test position became final
position.
Anemometer is mounted Due north of wind vane.
16
Other Tidbits - II
During the comparison tests on the BFN tower, 554
samples (6) blew through tower. Tower effects
on measurements can be examined.
Differences between 15-min. avgs (vane/anemometer
- ultrasonic)
WD (º) WS (mph) sT (º) Mean
0 2.1 -1 Range
-100/145 -0.4/7.6 -57/9 Std. Dev. 8
1.4 3
17
Conclusion
Care must be taken to verify that ultrasonic wind
sensors are suitable for operational
applications. Changes should be well
documented. Unexpected problems can occur.

• ULTRASONIC SENSORS HAVE PROVEN
• ACCEPTABLE FOR TVA AND REPRESENT
• A SIGNIFICANT IMPROVEMENT.
• Calibration tolerances are met consistently
• Less corrective maintenance has been necessary.